Recent trends suggest that inexpensive networked video sensing elements will
be pervasively deployed in our environment. Already, they are embedded in
devices such as computers and mobile phones and they are mounted in public
spaces such as malls and airports. In many ways, this trend could prove
beneficial to society, in that information collected by sensors could be
shared for the better good. Harnessing the power of these emergent sensory
environments will hinge on our ability to build applications capable of
gathering, interpreting and storing data from distributed sensors and to
provide scalable mechanisms for managing the networks and systems resources
that these applications consume.
The Sensorium Infrastructure and associated projects in
the Computer Science Department at Boston University aim to catalyze
fundamental advances in image and video computing, network protocols, and
resource management to deal with unique spatio-temporal constraints of
sensor networks in general and of video sensor networks in particular. When
fully acquired, the Sensorium research infrastructure will be composed of a
sensor network of video cameras spanning several rooms, networked processing
units, and a terabyte database, managed together to satisfy queries using
those generated by mobile users within this environment.
Sensorium infrastructure will enable a number of collaborative research
projects led by various research groups in the department. Information about
these projects is available in a 15-page
document describing in some depth the technical challenges and range of
problems to be tackled along the dimensions highlighted below.
The Image and
Video Computing Group will lead research on modeling, interpretation, and prediction of human
motion in video streams at multiple scales in space/time and at multiple
layers of detail.
The Web and
InterNetworking Group will lead research on the development of efficient location management, routing,
transport, and content distribution protocols for multi-resolution/scale
streaming sensory data networks. Also, it will work on efforts to
characterize and model network traffic and access patterns in mobile and wireless (video) sensor networks.
Operating Systems and Services Group will lead research on instrumentation of embedded real-time operating systems
to enable coordinated resource management, and on the development of middleware services for the management
of active and ad-hoc sensor networks.
Database Group will investigate issues in the indexing and mining of large spatio-temporal
non-textual sensory datasets, with a particular emphasis on mining of
human motions and activities.
Programming Languages and Systems Group will work on techniques to enhance code safety for embedded systems through the
use of type systems and run-time support, with emphasis on flow-oriented
and Applied Cryptography Group will focus on the development of algorithms for supporting security and
trust, and for protecting the confidentiality and integrity of data in
video sensor networks and repositories.